Page 263 - Ai Book - 10

NEURAL NETWORK

In the previous class, you have learnt about neural networks. As
you know, a neural network is a series of algorithms that depicts
the relationships in a set of data through a process that mimics
the way the human brain operates. In this sense, you can say
that neural networks refer to the systems of neurons which are
either organic or artificial in nature. It is also a neural network
which serves as a powerful technology in many computer vision
applications. Before learning convolutional neural networks in detail, you should demonstrate the knowledge
about neural networks by labelling the name of different layers used in the given image of a neural network.

Convolution Neural Network
As you know, convolution is the simple application of a filter to an input image that gives us an enhanced output
image. CNNs or ConvNets is the acronym of Convolution neural network, refers to a category of Neural Networks
that have proven very effective in various areas like image recognition and classification. ConvNets have been
successful in identifying faces, objects and traffic signs apart from powering vision in robots and self driving
cars. In short, CNNs are powerful deep Learning algorithms that assign weights and biases to various aspects/
objects available in the input image. The weights in the network are initialized to small random numbers ranging
from-1.0 to 1.0, or -0.5 to 0.5. Each unit has a bias associated with it. The biases are similarly initialized to small
random numbers.
In simple words, the convolutional neural network (CNN) is a multilayer, feed-forward neural network that
uses perceptrons for supervised learning and data analysis. It is used mainly with visual data, such as image
classification. These networks are basically designed to process data through multiple layers of arrays. As you
know, feature extraction is an important part of AI systems which is being performed by the neural networks.
The primary difference between CNN and any other ordinary neural network is that CNN takes input as a two-
dimensional array and operates directly on the images rather than focusing on feature extraction. The various
applications of CNNs are as follows:

u Product Recommendation Engine: Product
Recommendation Engine is a field where image
classification and object recognition can be easily
carried out by CNNs. For example, Amazon uses
CNN image recognition for suggestions in the
“you might also like” section. The basis of the
assumption is the user’s expressed behaviour. The
products itself are matched on visual criteria like
black high heels for the black dress.

u Medical Image Classification: Nowadays, medical practitioners
use CNNs medical image classification to detect the various kinds
of anomalies in an X-ray or MRI as higher precision is not possible
with human vision.

u Signature Verification: The verification of signature is an important application of CNNs. As you know, in the
banking and other financial industry, the recognition of personal signature becomes an extra validating and

137
137